7646

• UCREFRP Catalog Number: 7646
• Author: Gadomski, D. M. and J. A. Hall-Griswold
• Title: Predation by Northern Squawfish on Live and Dead Juvenile Chinook Salmon
• USFW Year: 1992
• USFW - Doc Type: Transactions of the American Fisheries Society
• Copyright Material: YES

682

NOTES

TABLE I.-Consumption of live and dead juvenile chinook salmon by northern squawfish in six laboratory
experiments. Trial results are pooled. Northern squawfish were offered either 50% dead and 50% live or 20% dead
and 80% Jive chinook salmon. Two experimental systems were used: 1,400-L circular tanks (CT) with 3 northern
squawfish per trial, and an 11,300-L raceway (RW) with 10 northern squawfish per trial. One set of trials was
conducted in complete darkness (CT-dark: <0.01 lux); an additional set of trials was in bright light only (CT-light:
215-270 lux). All other trials were in a natural photoperiod with both light and dark periods. Degrees of freedom
for pooled x2 = 1. Asterisks denote P < 0.05* or P < 0.0 I **.

Experimental conditions
Percen t Prey consumed
dead Duration Number Total Percent Pooled
salmon System (h) of trials number dead x2 I-{J
50 CT 24 18 332 61.8 18.3.. >0.95
50 RW 3 3 132 78.8 43.7.* >0.95
20 CT 24 12 118 35.6 17.9.* >0.95
20 RW 3 3 36 36.1 5.8* >0.75
20 CT -dark 4 22. 79 31.2 6.7** >0.75
20 CT-light 4 22 59 88.1 171.2** >0.95

· Twenty-seven CT -dark trials were conducted. but only 22 were used in analyses because more than 50% of the available prey
were consumed during 5 trials.

consumed (no selection) were I: I (50% dead prey)
and 1:4 (20% dead prey). The power (I - (J) of
the chi-square tests was computed from formulas
in Sokal and Rohlf ( I 981).

Results
More than 50% of prey were eaten in only five
trials (total N = 85). These trials, which were part
of the 4-h dark experiment, were excluded from
analysis.
All trials with 50% dead prey in the circular
tanks and raceway could be validly tested by a
heterogeneity chi-square analysis. During 14 of
the 18 circular tank trials, 50% or more of the
total number of prey consumed were dead. The
mean percentage of dead prey consumed was
63.6% (SO, 16.7; N = 18). Total (x2 = 53.8; df =
18) and pooled (x2 = 18.3; df = I; Table I) data
deviated significantly (P < 0.0 I) from the expect-
ed frequencies of live and dead prey consumed.
Between-trial heterogeneity (x2 = 35.5; df = 17)
was also significant (P < 0.0 I); however, this was
largely due to the relatively small number of prey
consumed per circular tank (mean, 18.4; SO, 5.0;
N = 18). The few prey consumed caused many
individual tests to have low power (range of I -
(j, 0.1-0.9) to detect a small percent change from
the expected frequency. In contrast, if data were
pooled, the power of the chi-square test exceeded
0.95 (Table I).
During the raceway trials with 50% dead prey,
the mean percentage of dead prey consumed was
78.5% (SO, 2.13; N = 3). Total (x2 = 44.0; df =
3) and pooled (X2 = 43.7; df = 1; Table 1) data
were significantly different (P < 0.01) from ex-

pected frequencies, but in contrast to the circular
tank replicates, heterogeneity was not significant
(x2 = 0.3; df= 2). Mean number of prey consumed
per raceway trial was 44.0 (SO, 9.6; N = 3). The
power of individual raceway tests to detect the
observed percent change from the expected fre-
quency exceeded 0.9.
In experiments with 20% dead prey, too few
prey were consumed per tank for unbiased chi-
square analyses of individual trial results. Mean
number of prey consumed per trial during the 24-h
circular tank experiment was 9.8 (SO, 3.1; N =
12); in the raceway, the mean number of prey
consumed was 12.0 (SO, 4.1; N = 3). Therefore,
trial results were pooled within each experimental
type. Pooled results differed significantly from ex-
pected frequencies (Table I).
Northern squawfish were much more selective
for dead chinook salmon during the light trials
than during the dark trials (Table I). During bright
light, when 20% dead prey were available, 88.1 %
of consumed prey were dead. Periodic counts of
uneaten prey during II trials in the light experi-
ment indicated that most dead prey were eaten
soon after they were placed in the tanks. Twenty-
two of 27 dead prey consumed (81.5%) were in-
gested within 0.5 h after trials began; 26 of27 dead
prey (96.3%) were ingested within 2 h. During the
same trials only two live prey were consumed,
both within the first 0.5 h.
Live prey were occasionally killed by northern
squawfish but not ingested. Pharyngeal tooth marks
on the sides of these individuals suggested that
northern squawfish had attempted to swallow these
prey tailfirst and then had regurgitated them. Re-